Magnesium accumula ion upon cyclin M4 silencing
ac i a es mic osomal iglyce ide ans e p o ein
imp o ing NASH
G aphical abs ac
Highligh s
CNNM4 ac s as a magnesium expo e in he li e . I s up egula ion in
NASH leads o ele a ed magnesium le els in se um.
Li e -specific CNNM4 a ge ing alle ia es s ea osis, inflamma ion,
and fib osis in p eclinical NASH models.
siRNA-media ed CNNM4 down egula ion p omo es hepa ic magne-
sium accumula ion and educes endoplasmic e iculum s ess.
Silencing CNNM4 enhances mic osomal iglyce ide ans e p o ein
ac i i y leading o VLDL assembly and sec e ion.
Au ho s
Jo ge Simón, Na oa Goikoe xea-
Usandizaga, Ma ina Se ano-Maciá,
., Te esa Ca doso Delgado, Luis
Al onso Ma ínez-C uz, Ma ía
Luz Ma ínez-Chan a
Co espondence
mlma [email protected]
(M.L. Ma ínez-Chan a )
Lay summa y
Cyclin M4 (CNNM4) is o e ex-
p essed in non-alcoholic s ea ohe-
pa i is (NASH) and p omo es he
expo o magnesium om he li e .
The li e -specific silencing o Cnnm4
amelio a es NASH by educing
endoplasmic e iculum s ess and
p omo ing he ac i i y o mic o-
somal iglyce ide ans e p o ein.
h ps://doi.o g/10.1016/j.jhep.2021.01.043
© 2021 Eu opean Associa ion o he S udy o he Li e . Published by Else ie B.V. This is an open access a icle unde he CC BY-NC-ND license
(h p://c ea i ecommons.o g/licenses/by-nc-nd/4.0/). J. Hepa ol. 2021, 75, 34–45
Resea ch A icle
NAFLD and Alcohol-Rela ed Li e Diseases
Magnesium accumula ion upon cyclin M4 silencing ac i a es
mic osomal iglyce ide ans e p o ein imp o ing NASH
Jo ge Simón
1,2
, Na oa Goikoe xea-Usandizaga
1
, Ma ina Se ano-Maciá
1
,
Da id Fe nández-Ramos
1,2,3
, Diego Sáenz de U u i
4
, Jessica J. G uskos
5
,
Pablo Fe nández-Tussy
1
, So ía Lachiondo-O ega
1
, I ene González-Recio
1
,
Rubén Rod íguez-Agudo
1
, Vi ginia Gu ié ez-de-Juan
1
, Begoña Rod íguez-I u e agoyena
1
,
Ma a Va ela-Rey
1,2
, Paula Gimenez-Masca ell
1
, Ma ía Me cado-Gomez
1
,
Bea iz Gómez-San os
4
, Ca men Fe nandez-Rod iguez
1
, Fe nando Lopi z-O soa
1,3
,
Maide Bizka guenaga
1,3
, Sibylle Dames
6
, U e Schaepe
6
, F anz Ma in
7,8
, Guadalupe Sabio
9
,
Paula I uzubie a
10,11
, Ja ie C espo
10,11
, Pa icia Aspichue a
2,4,12
, Ke an H.-Y. Chu
5
,
Daniela Buccella
5
, Césa Ma ín
13
, Te esa Ca doso Delgado
1
, Luis Al onso Ma ínez-C uz
1,†
,
Ma ía Luz Ma ínez-Chan a
1,2,
*
,†
1
Li e Disease Labo a o y, Cen e o Coope a i e Resea ch in Biosciences (CIC bioGUNE), Basque Resea ch and Technology Alliance (BRTA),
Bizkaia Technology Pa k, De io, Spain;
2
Cen o de In es igación Biomédica en Red de En e medades Hepá icas y Diges i as (CIBERehd), 48160,
Bizkaia, Spain;
3
P ecision Medicine and Me abolism Labo a o y, Cen e o Coope a i e Resea ch in Biosciences (CIC bioGUNE), Basque
Resea ch and Technology Alliance (BRTA), Bizkaia Technology Pa k, De io, Spain;
4
Depa men o Physiology, Facul y o Medicine and Nu sing,
Uni e si y o he Basque Coun y (UPV/EHU), Leioa, Bizkaia, Spain;
5
Depa men o Chemis y, New Yo k Uni e si y, New Yo k, NY, USA;
6
Silence The apeu ics GmbH, Be lin, Ge many;
7
Cen o Andaluz de Biología Molecula y Medicina Regene a i a-CABIMER, Uni e sidad Pablo
de Ola ide, Uni e sidad de Se illa, Consejo Supe io de In es igaciones Cien íficas (CSIC), Se ille, Spain;
8
Cen o de In es igación Biomédica en
Red de Diabe es y En e medades Me abólicas Asociadas (CIBERDEM), Mad id, Spain;
9
Fundación Cen o Nacional de In es igaciones
Ca dio ascula es Ca los III, Mad id, Spain;
10
Gas oen e ology and Hepa ology Depa men , Ma qués de Valdecilla Uni e si y Hospi al,
San ande , Spain;
11
Clinical and T ansla ional Diges i e Resea ch G oup, Ins i u o de In es igación Sani a ia Valdecilla (IDIVAL), San ande ,
Spain;
12
Bioc uces Heal h Resea ch Ins i u e, Ba akaldo, Bizkaia, Spain;
13
Ins i u o Biofisika (UPV/EHU, CSIC) and Depa amen o de
Bioquímica, Uni e sidad del País Vasco, Bilbao, Spain
Backg ound & Aims: Pe u ba ions o in acellula magnesium
(Mg
2+
) homeos asis ha e implica ions o cell physiology. The
cyclin M amily, CNNM, pe o m key unc ions in he anspo o
Mg
2+
ac oss cell memb anes. He ein, we aimed o elucida e he
ole o CNNM4 in he de elopmen o non-alcoholic s ea ohe-
pa i is (NASH).
Me hods: Se um Mg
2+
le els and hepa ic CNNM4 exp ession
we e cha ac e ised in clinical samples. P ima y hepa ocy es we e
cul u ed unde me hionine and choline dep i a ion. A 0.1%
me hionine and choline-deficien die , o a choline-deficien
high- a die we e used o induce NASH in ou in i o oden
models. Cnnm4 was silenced using siRNA, in i o wi h Dha -
maFECT and in i o wi h In i o ec amine
®
o conjuga ed o N-
ace ylgalac osamine.
Resul s: Pa ien s wi h NASH showed hepa ic CNNM4 o e -
exp ession and dys egula ed Mg
2+
le els in he se um. Cnnm4
silencing amelio a ed hepa ic lipid accumula ion, inflamma ion
and fib osis in he oden NASH models. Mechanis ically, CNNM4
knockdown in hepa ocy es induced cellula Mg
2+
accumula ion,
educed endoplasmic e iculum s ess, and inc eased mic o-
somal iglyce ide ans e ac i i y, which p omo ed hepa ic lipid
clea ance by inc easing he sec e ion o VLDLs.
Conclusions: CNNM4 is o e exp essed in pa ien s wi h NASH
and is esponsible o dys egula ed Mg
2+
anspo . Hepa ic
CNNM4 is a p omising he apeu ic a ge o he ea men o
NASH.
Lay summa y: Cyclin M4 (CNNM4) is o e exp essed in non-
alcoholic s ea ohepa i is (NASH) and p omo es he expo o
magnesium om he li e . The li e -specific silencing o Cnnm4
amelio a es NASH by educing endoplasmic e iculum s ess and
p omo ing he ac i i y o mic osomal iglyce ide ans e p o ein.
© 2021 Eu opean Associa ion o he S udy o he Li e . Published by
Else ie B.V. This is an open access a icle unde he CC BY-NC-ND
license (h p://c ea i ecommons.o g/licenses/by-nc-nd/4.0/).
In oduc ion
Die a y imbalances, such as a low in ake o magnesium, a e
ecognised as he oo cause o diseases. Al hough daily mag-
nesium ecommended in ake is 300–400 mg/day, he e is a
g owing conce n abou dec eased in ake in las decade, a ib-
u able o changes in die a y habi s and ood p ocessing.
1,2
I s
ionic o m, Mg
2+
, is he mos abundan di alen ca ion in he cell
and i is equi ed as a co ac o o o e 300 enzyma ic eac ions.
3
Keywo ds: Non-alcoholic s ea ohepa i is; NASH; Cyclin M4; CNNM4; Magnesium;
The apy; siRNA; Endoplasmic e iculum s ess; Mic osomal iglyce ide ans e
p o ein; MTP.
Recei ed 20 Ap il 2020; ecei ed in e ised o m 18 Janua y 2021; accep ed 19 Janua y
2021; a ailable online 9 Feb ua y 2021
*Co esponding au ho . Add ess: Li e Disease Labo a o y, Building 801A,
Technologic Pa k o Biscay, De io (Biscay), Spain. Tel.: +34-944-061-318; Fax: +34-
944-061-301.
E-mail add ess: [email p o ec ed] (M.L. Ma ínez-Chan a ).
†
Senio au ho ship.
h ps://doi.o g/10.1016/j.jhep.2021.01.043
Jou nal o Hepa ology 2021 ol. 75 j34–45
Resea ch A icle
NAFLD and Alcohol-Rela ed Li e Diseases
An in ica e ne wo k o Mg
2+
anspo e s pa icipa e in i s up-
ake and exc e ion, allowing i s flux ac oss cellula memb anes,
hus impac ing Mg
2+
homeos asis and dis ibu ion.
1
Mg
2+
supplemen a ion educes mo ali y om hepa ic com-
plica ions a ising om alcohol in ake o s ea osis.
4
Hypo-
magnesaemia is p esen in se e al como bidi ies o non-
alcoholic a y li e disease (NAFLD) such as insulin esis ance
and ype 2 diabe es,
5
ca dio ascula complica ions,
6
and
obesi y.
7
Mo eo e , deficiencies in Mg
2+
a e ela ed o inflam-
ma o y esponses, mi ochond ial dys unc ion, and dec eased
ac i i y o he an ioxidan sys em, all ea u es o li e diseases.
8
The e m NAFLD encompasses a g oup o pa hologies cha -
ac e ised by ch onici y wi h se ial p og ession om s ea osis o
non-alcoholic s ea ohepa i is (NASH) and ci hosis.
9
Such p o-
g ession is media ed by he inc eased p oduc ion o eac i e
oxygen species (ROS), lipo oxici y, mi ochond ial dys unc ion,
and he de elopmen o endoplasmic e iculum (ER) s ess.
10
Rema kably, ER s ess has been p oposed o be an impo an
mechanism because i dis up s calcium (Ca
2+
) anspo h ough
ATPases
11
and leads o p o ein mis olding.
12
I has been widely
cha ac e ised in bo h NAFLD and NASH pa ien s, oge he wi h
impai ed VLDL assembly and sec e ion.
13,14
NAFLD is a global heal h p oblem wi h an es ima ed p e a-
lence o a ound 25% among he adul popula ion, being expec ed
o inc ease.
9
Changes in li es yle habi s a e he mos common
ecommenda ions o he clinical managemen o NAFLD, bu i is
di ficul o achie e he long- e m compliance o pa ien s, making
pha macological al e na i es a ac i e.
Among he a ious magnesio opic p o eins, he cyclin M
(CNNM) amily and hei Ba eman-module-media ed in e ac ions
wi h phospha ases o egene a ing li e (PRLs) a e o pa icula
in e es .
15
PRLs a e implica ed in umou p og ession and me as-
asis,
16,17
including li e cance , wi h Mg
2+
pe u ba ions epo ed
in such.
18
Al hough he CNNMs eme ged as in e ac ing molecula
pa ne s o PRLs,
17
a new pe spec i e p oposes a ge ing CNNMs o
modula e Mg
2+
homeos asis mo e specifically.
19
The unc ions o
hepa ic CNNMs in his ega d a e unknown, and hei significance
as pha macological a ge s emains unde explo ed.
In he p esen s udy, we demons a ed he ele ance o he
di e en ial exp ession o CNNM4 in NASH de elopmen , showing
i s unc ional ole in anspo ing hepa ic Mg
2+
.Mo eo e ,we
iden ified CNNM4 as a po en ial a ge o NASH ea men .
Pa ien s and me hods
Pa ien s
Measu emen s o se um Mg
2+
and hepa ic CNNM1-4 mRNA, and
immunohis ochemical CNNM4 analysis we e pe o med in
di e en coho s ec ui ed a he Hospi al Ma qués de Valdecilla,
San ande , Spain (Fig. S1A and B).
Animal main enance and p eclinical s udies
Mice we e main ained wi h ad libi um access o wa e and a
choline-deficien die wi h 0.1% me hionine (0.1% MCDD) o a
choline-deficien , high- a die (CD-HFD). A con ol g oup was
main ained on a egula die (SC die ).
T ea men o p ima y mouse hepa ocy es and THLE-2 cells
Upon a achmen , isola ed mouse p ima y hepa ocy es we e
ans ec ed by o e nigh incuba ion wi h siRNA using Dha ma-
FECT 1 o a CNNM4 exp ession plasmid using je PRIME
®
. Cells
we e main ained and incuba ed o an addi ional 24 h unde
di e en condi ions.
Cnnm4 silencing in i o
Mice ed a 0.1% MCDD o 2 weeks o a CD-HFD o 3 weeks we e
epea edly adminis e ed an siRNA using In i o ec amine
®
3.0
Reagen h ough ail ein injec ion. Mice we e sac ificed a e 4
weeks on 0.1% MCDD and 6 weeks on CD-HFD, espec i ely. Fo
Cnnm4 GalNAc-siRNA deli e y, mice ed o 3 weeks on 0.1%
MCDD we e ea ed once, subcu aneously, and sac ificed a e a
o al o 6 weeks on he die . Samples o li e , whi e adipose
issue (WAT), and se um we e collec ed.
S a is ical analysis
All he expe imen s we e pe o med a leas in iplica e, wi h n = 3
(in i o)andn=4(in i o). The da a a e exp essed as mean ± SEM
and ep esen he old change s. con ol mean alue when indi-
ca ed. S a is ical significance was de e mined using P ism 8
(G aphPad So wa e, San Diego, CA, USA). G oups we e compa ed
by 1-way analysis o a iance (ANOVA) ollowed by pos hoc Bon-
e oni es s ( o 3 o mo e g oups) o he S uden es ( o 2
g oups).
Resul s
CNNM4 is o e exp essed in clinical and p eclinical NASH
To in es iga e Mg
2+
dys egula ion in NASH, we de e mined Mg
2+
le els in se a om a coho o 50 age- and BMI-ma ched pa ien s
(Fig. S1A), obse ing inc eased le els o he ca ion in NASH pa-
ien s (Fig. 1A). Mg
2+
le els did no co ela e wi h o he NASH
bioma ke s (da a no shown).
Conside ing he magnesio opic ole o he CNNM amily,
19
we
sough o assess whe he dys egula ed CNNM exp ession could be
he cause o he ele a ed se um Mg
2+
. The hepa ic mRNA le els o
each CNNM we e assessed in ano he coho o 40 pa ien s (Fig.
S1B), showing a significan o e exp ession o CNNM1 and CNNM4
in NASH pa ien s (Fig.1B). In addi ion, an inc ease in Cnnm4 mRNA
exp ession was obse ed in p ima y hepa ocy es ea ed wi h
me hionine and choline-deficien (MCD) medium (Fig. 1C), an in
i o model ha displays ea u es o NASH such as ROS o e -
p oduc ion and lipid accumula ion.
20,21
Simila ly, hepa ic CNNM4 o e exp ession in NASH was
confi med by immunohis ochemical p o ein de e mina ion in
clinical samples (Fig. 1D) and animal models –mice ed a 0.1%
MCDD (Fig. 1E) and CD-HFD (Fig. 1F). These animal models also
showed ele a ed Cnnm4 mRNA, as did p ima y hepa ocy es (Fig.
S1C and D). By con as , Cnnm1 o e exp ession was obse ed in
nei he he in i o no in i o NASH models. To e alua e p o ein
s abili y, human THLE-2 cells we e ea ed wi h MLN4924 o
inhibi NEDDyla ion, which p e en s p o easome-media ed
deg ada ion,
22
esul ing in dec eased CNNM4 exp ession (Fig.
S1E). The exp ession o o he Mg
2+
anspo e s was de e -
mined in animal models (Fig. S1F) and clinical samples (Fig. S1G),
showing ha CNNM4 is unique among Mg
2+
anspo e s in i s
up egula ion in all condi ions.
Ta ge ing CNMM4 amelio a es NASH
The possible ole o CNNM4 in NASH de elopmen was exam-
ined by in i o sc eening by specifically silencing each Cnnm in
p ima y hepa ocy es using small in e e ing RNA (siRNA). Ta -
ge ing Cnnm4 (siCnnm4) alone e ec i ely educed MCD-induced
lipid accumula ion (Fig. 2A). The exp ession o each Cnnm and P l
Jou nal o Hepa ology 2021 ol. 75 j34–45 35
F
0
1
2
3
4
5
SC die
Cnnm4 a ea (a.u.)
2 weeks
***
4 weeks
****
CD-HFD
SC die 3 weeks 6 weeks
CD-HFD
CNNM4
E
0
5
10
15
Cnnm4 a ea (a.u.)
SC die
*
2 weeks
****
4 weeks
0.1% MCDD
SC die 2 weeks 4 weeks
0.1% MCDD
CNNM4
D
0
5
10
15
20
25
Cnnm4 a ea (a.u.)
Heal hy
*
S ea osis
***
NASH
Heal hy NASHS ea osis
CNNM4
C
0.0
0.5
1.0
1.5
2.0
C l
MCD
*
Rela i emRNA
exp ession ( old-change)
Cnnm1 Cnnm2 Cnnm3 Cnnm4
B
0
1
2
3
4
5
Rela i e mRNA
exp ession ( old-change)
**** *** Heal hy (n = 5)
S ea osis (n = 20)
NASH (n = 15)
CNMM1 CNNM2 CNNM3 CNNM4
Heal hy (n = 18)
NASH (n = 18)
S ea osis (n = 14)
0
1
2
3
4*
Magnesiumin
se um (mg/dl)
A
Fig. 1. CNNM4 is o e exp essed in NASH pa ien s and p eclinical animal models. (A) Magnesium de e mina ion in se um and (B) hepa ic le els o CNNM1-4
mRNA in di e en coho s o heal hy indi iduals, pa ien s wi h s ea osis and pa ien s wi h NASH. (C) Le els o Cnnm1-4 mRNA in p ima y hepa ocy es s imula ed
wi h MCD medium. (D) Li e immunohis ochemical s aining and quan ifica ion o CNNM4 in a coho o heal hy indi iduals, pa ien s wi h s ea osis and pa ien s
wi h NASH, as well as mice ed (E) a 0.1% MCDD o (F) a CD-HFD. Scale ba co esponds o 100
l
m. *p<0.05, ***p<0.001, and ****p<0.0001 s. heal hy/c l/SC die .
CD-HFD, choline-deficien high- a die ; MCD, me hionine and choline deficien ; MCDD, MCD die ; NASH, non-alcoholic s ea ohepa i is; SC die , egula die .
B
0
3
6
9
12
CNNM4 a ea (a.u.)
*
#20
30
40
0.1% MCDD
siC l
SC die
0.1% MCDD
siCnnm4
siCnnm4
siC l
SC die
0
10
#
***
Su an ed a ea (a.u.)
CNNM4SUDAN RED
0.1% MCDD
siCnnm4siC lSC die
CNNM4 a ea (a.u.)
0
1
2
3
4
**
##
5
10
15
CD-HFD
siC l
SC die
CD-HFD
siC l
SC die
Su an ed a ea (a.u.)
0
*
#
CD-HFD
CNNM4SUDAN RED
siCnnm4siC lSC die
C
siCnnm4
siCnnm4
A
0
1
2
3
4
5
MCD
***
siCnnm1
***
siCnnm2
***
siCnnm3
###
siCnnm4
C l
***
siC l
In ensi y o lipid bodies/
cells ( old-change)
BODIPY
C l MCD + siC l
MCD + siCnnm2 MCD + siCnnm3
MCD + siCnnm1
MCD + siCnnm4
Fig. 2. Ta ge ing CNNM4 educes lipid accumula ion in in i o and in i o NASH models. (A) BODIPY s aining mic og aphs and quan ifica ion in p ima y
hepa ocy es ans ec ed wi h siRNA agains Cnnm1-4 (siCnnm1-4) and s imula ed wi h MCD o 24 h. Mic og aphs o CNNM4 and Sudan Red s aining and
espec i e quan ifica ion in mice ed (B) a 0.1% MCD die (0.1% MCDD) o (C) a CD-HFD, and ea ed wi h Cnnm4 siRNA (siCnnm4) o an un ela ed con ol (siC l).
Scale ba co esponds o 50
l
m. *p<0.05, **p<0.01, and ***p<0.001 s. C l/SC die ;
#
p<0.05,
##
p<0.01, and
###
p<0.001 s. MCD + siC l/0.1% MCDD + siC l/CD-
HFD + siC l. CD-HFD, choline-deficien high- a die ; MCD, me hionine and choline-deficien ; MCDD, MCD die ; NASH, non-alcoholic s ea osis; SC die , egula
die ; siRNA, small in e e ing RNA.
36 Jou nal o Hepa ology 2021 ol. 75 j34–45
Resea ch A icle NAFLD and Alcohol-Rela ed Li e Diseases
mRNA was e alua ed, confi ming he specific and e ec i e
silencing o Cnnm (Fig. S2A) and elimina ing he possibili y o P l
egula ion when a ge ing Cnnm4 (Fig. S2B). Addi ionally,
CNNM4 silencing in THLE-2 cells also educed MCD-induced lipid
accumula ion (Fig. S2C and D).
Taking in o conside a ion he CNNM4 o e exp ession
obse ed in NASH, and he siCnnm4-de i ed educ ion o lipid
accumula ion, he he apeu ic po en ial o silencing Cnnm4 in
i o was explo ed. Mice we e ed a 0.1% MCDD o induce s ea-
osis de elopmen , inc eased oxida i e s ess, inflamma ion, and
fib osis in a sho pe iod o ime.
23
A 2 weeks, by which s ea-
osis
21
and CNNM4 o e exp ession a e obse able (Fig. 1E), 0.1%
MCDD- ed mice we e sepa a ed in o 2 g oups and ea ed
epea edly wi h ei he an siRNA agains Cnnm4 (0.1% MCDD +
siCnnm4) o an un ela ed con ol (0.1% MCDD + siC l), using
In i o ec amine 3.0
®
. The mice we e sac ificed a he ou h
week. In e es ingly, he mice ed he 0.1% MCDD showed hepa ic
CNNM4 o e exp ession, whe eas Cnnm4 silencing educed die -
induced s ea osis (Fig. 2B and Fig. S2E). Alpha-smoo h muscle
ac in (
a
SMA) s aining e ealed ha fib osis, ano he NASH
hallma k, was induced by 0.1% MCDD and a enua ed by siCnnm4
(Fig. S2F). Fu he mo e, he ea men also educed se um
alanine amino ans e ase (ALT) le els, a ma ke o li e damage
(Fig. S2G).
In a di e en NASH mu ine model, mice we e ed a CD-HFD,
as his leads o a pa e n o pa hology mo e simila o ha
obse ed in humans.
24
Al hough s udies using his model end o
be pe o med o longe han 6 weeks,
24
ou g oup p e iously
showed ha mice de eloped ea ly NASH pheno ypes comp ising
weigh gain, s ea osis, and inflamma ion as ea ly as his poin in
ime.
21
A e 3 weeks on he die , a which poin CNNM4 was
al eady o e exp essed (Fig. 1F), he CD-HFD- ed g oup was
di ided and ea ed wi h siCnnm4 (CD-HFD + siCnnm4) o siC l
(CD-HFD + siC l). They we e sac ificed a e 6 weeks and,
simila ly o in he p e ious esul s, he mice ed he CD-HFD
showed hepa ic CNNM4 o e exp ession, while he siRNA he -
apy educed s ea osis (Fig. 2C and Fig. S2H) and fib osis de el-
opmen (Fig. S2I). Se um ALT le els emained unal e ed (Fig. S2J).
G
Heal hy li e
O he Mg2+
expo e s
O he Mg2+
expo e s
O he Mg2 +
expo e s
NASH li e
Mg2+
Mg
2+
Mg2+
Mg2+
Mg2+
Mg2+
siCnnm4
siCnnm4- ea ed li e
Cnnm4
Cnnm4
F
E
2.0
Magnesium in
cul u e medium (mg/dl)
0.0
0.5
1.0
1.5
siCnnm1
siCnnm2
siCnnm3
siC l
0.0
0.2
0.4
0.6
siCnnm1
siCnnm2
siCnnm3
siC l
Non- a ge ed
luo escence a io (a.u.)
0.0
0.2
0.4
0.6
siCnnm1
siCnnm2
siCnnm3
siC l
Mi ochond ial- a ge ed
luo escence a io (a.u.)
D
0.0
0.3
0.6
0.9
1.2
siC l
siCnnm4
Magnesium in
cul u e medium (mg/dl)
#
C
Mi o acke REDMe ge
siCnnm4
siC l
Non- a ge ed
labeling
Mi o acke RED
Me ge
siCnnm4
siC l
Mi ochond ial
a ge ed labeling
0.00
0.15
0.30
0.45
0.60
Non- a ge ed
luo escence a io (a.u.)
#
siC l
siCnnm4
0.00
0.15
0.30
0.45
Mi ochond ial- a ge ed
luo escence a io (a.u.)
siC l
siCnnm4
##
B
MagS
Non a ge ed
MagS-TPP
Mi ochond ia a ge ed
A
0.0
0.3
0.6
0.9
1.2
0.1% MCDD
siCnnm4
#
siC l
SC die
Magnesium in
se um (mg/dl)
0.0
0.5
1.0
1.5
2.0
CD-HFD
siCnnm4
#
**
siC l
SC die
Magnesium in
se um (mg/dl)
Fig. 3. Magnesium dis ibu ion a e specific silencing o Cnnm4.(A) Magnesium in se um om mice ed a 0.1% MCDD o CD-HFD wi h specificCnnm4
silencing (siCnnm4) compa ed wi h non- ea ed (siC l) mice. (B) Biochemical s uc u e o non- a ge ed MagS and MagS-TPP p obes. (C) Mic og aphs and ela i e
in acellula magnesium le els and (D) ex acellula magnesium le els in p ima y hepa ocy es ea ed wi h an siRNA agains Cnnm4 (siCnnm4). Scale ba co -
esponds o 100
l
m. (E) In acellula magnesium and (F) ex acellula magnesium le els in p ima y hepa ocy es ea ed wi h siCnnm1-3. (G) Schema ic ep-
esen a ion o CNNM4-dependen magnesium fluc ua ions in li e and in ci cula ion **p<0.01 s. SC die ;
#
p<0.05, and
##
p<0.01 s. 0.1% MCDD + siC l/CD-HFD
+ siC l/siC l. CD-HFD, choline-deficien high- a die ; CNNM4, cyclin M4; MagS, magnesium-specific; MagS-TPP, mi ochond ion- a ge ed magnesium-specific;
MCDD, me hionine and choline deficien die ; SC die , egula die ; siRNA, small-in e e ing RNA.
Jou nal o Hepa ology 2021 ol. 75 j34–45 37
CNNM4 ac s in he li e as a magnesium expo e
Despi e he magnesio opic ole o CNNM4 in kidney epi helia,
25
i s physiological unc ion in he li e emains la gely unknown.
The possible ela ionship be ween inc eased se um Mg
2+
le els
and hepa ic CNNM4 o e exp ession (Fig. 1A and D) led us o
an icipa e a modula ion o he ca ion in he s udied NASH animal
models. The measu emen o Mg
2+
in he se um e ealed an
inc ease in CD-HFD- ed mice and a dec ease unde siCnnm4
ea men in bo h he CD-HFD and 0.1% MCDD models (Fig. 3A).
To confi m he hypo hesis o CNNM4 being a hepa ic Mg
2+
expo e , CNNM4- ela ed Mg
2+
flux in hepa ocy es was assessed
in i o. Fluo escen s aining wi h he a iome ic p obe MagS
was used o es ima e he ela i e le els o cy osolic ee Mg
2+
in
li e cells ea ed wi h siCnnm4. MagS is an analogue o Mag-
FURA-2, de eloped by he Buccela g oup,
26
which displays
simila me al ecogni ion bu enhanced op ical p ope ies.
Fu he mo e, aking in o conside a ion he key ole o mi o-
chond ia in hepa ocy e unc ion,
27,28
we sough o explo e he
mi ochond ial Mg
2+
le els in he a o emen ioned siCnnm4 con-
di ions. Fo his pu pose, we de eloped MagS-TPP, a a ge ed
a ian unc ionalised wi h a phosphonium g oup o deli e y o
he mi ochond ial ma ix (Fig. 3B). The new dye exhibi s a
selec i i y p ofile simila o MagS and binds o Mg
2+
wi h an
appa en dissocia ion cons an sui able o he de ec ion o
ypical in acellula concen a ions o ee Mg
2+
(Fig. S3).
Mi ochond ion- and non- a ge ed fluo escen indica o s we e
applied, e ealing an inc ease in bo h cy osolic and mi ochon-
d ial ee Mg
2+
le els in cells wi h educed Cnnm4 exp ession
(Fig 3C), combined wi h a dec ease in he ca ion in he
ex acellula medium (Fig 3D). The silencing o o he Cnnm
amily membe s did no lead o significan al e a ions in in a- o
ex acellula Mg
2+
(Fig. 3E and F), hus elimina ing hei possible
con ibu ion o Mg
2+
homeos asis in he hepa ocy e. These e-
sul s suppo he no ion ha he obse ed dec ease in se um
Mg
2+
o mice ea ed wi h siCnnm4 could be a consequence o i s
accumula ion in he li e (Fig. 3G).
CNNM4-media ed magnesium accumula ion educes lipid
con en
The associa ion be ween CNNM4 and magnesium e flux
p omp ed us o cha ac e ise he con ibu ion o Mg
2+
homeo-
s asis o NASH de elopmen . The ela i e con en o he ca ion in
p ima y hepa ocy es unde MCD condi ions was de e mined,
showing an MCD-induced dec ease in Mg
2+
and an inc ease
upon silencing Cnnm4 (Fig. 4A). Taking he la e in o conside -
a ion, a possible in e se ela ionship be ween hepa ic Mg
2+
and
lipid con en was in es iga ed. Magnesium deple ion in he cell
medium (0 mM Mg
2+
) esul ed in an inc eased hepa ocy e lipid
con en , while silencing Cnnm4 educed his e ec (Fig. 4B). In
line wi h he esul s o a clinical ial add essing he beneficial
p ope ies o magnesium,
4
Mg
2+
supplemen a ion (5 mM Mg
2+
)
e e ed MCD-induced lipid accumula ion in hepa ocy es
wi hou e ec s unde no mal condi ions (1 mM Mg
2+
)(Fig. S4A).
Nei he Mg
2+
deple ion no supplemen a ion a ec ed Cnnm4
exp ession (da a no shown).
To elimina e he possibili y o o he Mg
2+
anspo e s
con ibu ing o siCnnm4-induced lipid educ ion, hei exp es-
sion was cha ac e ised unde a ious condi ions. We obse ed an
E
In ensi y o lipid bodies/
cell ( old-change)
0
1
2
3
4
5
Emp y
ec o
CNNM4
ec o
1 mM Mg2+
5 mM Mg2+
***
1 mM Mg2+
***
5 mM Mg2+
Emp y ec o CNNM4 ec o
BODIPY
1 mM Mg2+ 5 mM Mg2+ 1 mM Mg2+ 5 mM Mg2+
D
0
20
40
60
80
Emp y ec o
CNNM4 ec o
***
Rela i e CNNM4 mRNA
exp ession ( old-change)
C
0.0
0.1
0.2
0.3
0.4
0.5
Non- a ge ed
luo escence a io (a.u.)
Emp y ec o
**
CNNM4 ec o
0.0
0.1
0.2
0.3
0.4
0.5
**
Mi ochond ial- a ge ed
luo escence a io (a.u.)
Emp y ec o
CNNM4 ec o
0
1
2
3
4
5
6
7
***
##
****
B
BODIPY
siCnnm4siC lSC die
0 mM Mg2+ media
0 mM Mg2+
siCnnm4
siC l
C l
In ensi y o lipid bodies/
cell ( old-change)
A
*
*
###
MCD
siC l
C l
0.0
0.2
0.4
0.6
Non- a ge ed
luo escence a io (a.u.)
*
###
*
MCD
siC l
C l
0.0
0.2
0.4
0.6
Mi ochond ial- a ge ed
luo escence a io (a.u.)
siCnnm4
siCnnm4
Fig. 4. Modula ion o in acellula lipid con en by CNNM4. (A) Rela i e in acellula magnesium de e mina ion by MagS and MagS-TPP s aining in p ima y
hepa ocy es unde MCD s imula ion and ea ed wi h a Cnnm4 siRNA (siCnnm4). (B) BODIPY s aining mic og aphs and quan ifica ion in mu ine hepa ocy es
main ained o 24 h unde magnesium deple ion (0 mM Mg
2+
) and ans ec ed wi h siCnnm4. (C) Rela i e in acellula magnesium de e mina ion and (D) Cnnm4
mRNA le els in p ima y hepa ocy es ans ec ed o 6 h wi h an emp y o CNNM4 ec o . (E) BODIPY s aining mic og aphs and espec i e quan ifica ion in
isola ed mouse hepa ocy es ans ec ed o 6 h wi h an emp y/CNNM4 exp ession ec o and s imula ed o 24 h wi h magnesium (5 mM Mg
2+
) s. a con ol
g oup (1 mM Mg
2+
). Scale ba co esponds o 100
l
m. *p<0.05, **p<0.01, ***p<0.001, and ****p<0.0001 s. C l/Emp y ec o ;
##
p<0.01 and
###
p<0.001 s. MCD
+ siC l/0 mM Mg
2+
+ siC l. CD-HFD, choline-deficien high- a die ; CNNM4, cyclin M4; MagS, magnesium-specific; MagS-TPP, mi ochond ion- a ge ed mag-
nesium-specific; MCD, me hionine and choline deficien ; siRNA, small in e e ing RNA.
38 Jou nal o Hepa ology 2021 ol. 75 j34–45
Resea ch A icle NAFLD and Alcohol-Rela ed Li e Diseases
inc ease in he exp ession o T ansien Recep o P o ein Melas a in
6(T pm6)(Fig. S4B), sugges ing ha he lipid educ ion may also
be he esul o inc eased Mg
2+
en y. As he e is no comme cial
inhibi o o TRPM6, i s iso o m 7 was inhibi ed wi h 2-
aminoe hyl diphenyl bo ina e (2-APB)
29
o u he in es iga e
he con ibu ion o Mg
2+
en y. We obse ed ha he lipid
accumula ion induced by inhibi ing Mg
2+
impo was no malised
by siCnnm4 (Fig. S4C). Finally, he unc ional ela ionship be-
ween cellula Mg
2+
le els and s ea osis was in es iga ed by
o e exp essing CNNM4 ia ansien ans ec ion, which educed
Mg
2+
le els (Fig. 4C and D) and aised lipid con en s in hepa o-
cy es (Fig. 4E). Rema kably, Mg
2+
supplemen a ion did no
a enua e CNNM4-induced lipid accumula ion (Fig. 4D and E).
Endoplasmic e iculum and oxida i e s ess a e educed by
silencing cyclin M4
Conside ing ha ROS p oduc ion and inflamma ion a e majo
d i e s o NASH p og ession,
10,28
mi ochond ial ROS we e
assessed in p ima y hepa ocy es, showing a educ ion in MCD-
induced ROS p oduc ion upon silencing Cnnm4 (Fig. 5A). The
de elopmen o ER s ess is also conside ed a ‘second hi ’, linked
o oxida i e s ess
30
and wi h an exis ing Mg
2+
flux be ween he
ER and mi ochond ia.
31
Thus, we hypo hesised ha ER in eg i y
migh be a ec ed by CNNM4-induced Mg
2+
fluc ua ions in NASH.
The e o e, cy osolic calcium (Ca
2+
), an ER-s ess indica o , was
quan ified in p ima y hepa ocy es. The inc eased [Ca
2+
]
cy osol
obse ed unde MCD s imula ion and siRNA-de i ed a enua ion
(Fig. 5B), oge he wi h he pa ial co-localisa ion o Mg
2+
and ER
(Fig. S5A), sugges ed a p o ec i e e ec o silencing Cnnm4, no
only o mi ochond ia, bu also o he ER. This was u he
add essed by cha ac e ising he elease o Ca
2+
om he ER by
s imula ing p ima y hepa ocy es wi h ATP, epo ed o p omo e
he P2Y ecep o -media ed elease o Ca
2+
in o he cy osol.
32
In e es ingly, an MCD-induced dec eased capaci y and hen
no malisa ion upon silencing Cnnm4 we e obse ed (Fig. 5C),
while ER labelling was dec eased unde he MCD and eco e ed
by Cnnm4 siRNA (Fig. 5D). The ela ionship be ween ER s ess
and hepa ic lipid con en was elucida ed by ea ing p ima y
G
0
2
4
5
10
15
0
1
2
3
4
Rela i e p o ein exp ession
no malized ( old-change)
p-eIF2
α/
elF2α
**
*
GRP78/
GAPDH
*
#
***
###
****
XBP1s/
GAPDH
**
###
****
*
p-AMPK/
AMPK
p-ERK 1/2/
ERK 1/2
p-S6/
S6
p-eIF2
α/
elF2α
GRP78/
GAPDH
XBP1s/
GAPDH
p-AMPK/
AMPK
p-ERK 1/2/
ERK 1/2
p-S6/
S6
*
#
##
CD-HFD + siC l
SC die
CD-HFD + si
Cnnm4
0.1%MCDD + siC l
SC die
0.1%MCDD + si
Cnnm4
F
0
20
40
60
DHE s aining (a.u.)
##
0.1%MCDD
0.1%MCDD
DHE
siCnnm4siC l
siCnnm4
siC l
E
0
2
4
6
8
Vehicle
Tunicamycin 0.5 μg/ml
Tunicamycin 1 μg/ml
**
##
#*
#
MCD
siCnnm4
siC l
C l
In ensi y o lipid bodies/
cells ( old-change)
ΠΠ
Π
D
0.0
0.5
1.0
1.5
2.0
*
#
MCD
siCnnm4
siC l
C l
Rela i e ER s aining
( old-change)
ER- acke
C l
MCD + siCnnm4MCD + siC l
A
C
0.0
0.5
1.0
1.5
2.0
2.5
##
MCD
siCnnm4
siC l
C l
ER Ca2+ elease capaci y
( old-change)
B
0
50
100
150
200
250
***
###
MCD
siCnnm4
siC l
C l
Cy osolic Ca2+ (nM)
0.0
0.5
1.0
1.5
2.0
***
###
MCD
siCnnm4
siC l
C l
Rela i e mi ochond ial
ROS ( old-change)
Fig. 5. SpecificCnnm4 silencing educes oxida i e and ER s ess in in i o and in i o NASH models. (A) Rela i e mi ochond ial ROS le els, (B) cy osolic Ca
2+
le els, (C) Ca
2+
elease capaci y o ER o e ime and (D) mic og aphs o ER- acke ed s aining and quan ifica ion in p ima y hepa ocy es cul u ed o 24 h in a
MCD medium and ans ec ed wi h a Cnnm4 (siCnnm4) o con ol (siC l) siRNA. Scale ba co esponds o 100
l
m. (E) BODIPY de e mina ion in p ima y he-
pa ocy es ea ed o 24 h wi h unicamycin, MCD, o C l medium and ea ed wi h siCnnm4 o siC l. (F) Rep esen a i e mic og aphs (scale ba co esponds o 50
l
m) and DHE s aining o li e om mice ed a 0.1% MCDD and ea ed wi h siCnnm4 o siC l. (G) Wes e n blo analysis o binding immunoglobulin p o ein (BIP/
GRP78), x-box binding p o ein 1 iso o m s (XBP1s), phospho
Se 51
-euka yo ic ini ia ion ac o 2
a
(eIF2
a
), phospho
Th 172
-AMP dependen p o ein kinase (AMPK),
phospho
Th 202/Thy 204
-ERK1/2, and phospho
Se 235/236
-S6 ibosomal p o ein (S6). Mouse li e s om di e en g oups we e compa ed: heal hy (SC die ), 0.1% MCDD,
o CD-HFD ea ed wi h siCnnm4 o siC l. *p<0.05, **p<0.01, ***p<0.001, and ****p<0.0001 s. C l/C l + ehicle/SC die ;
#
p<0.05,
##
p<0.01, and
###
p<0.001 s.
MCD + siC l/0.1% MCDD + siC l/MCD + ehicle + siC l/CDHFD + siC l;
p
p<0.05 and
pp
p<0.01 o ehicle s. unicamycin. CD-HFD, choline-deficien high- a
die ; CNNM4, cyclin M4; DHE, dihyd oxye hyl; ER, endoplasmic e iculum; MCD, me hionine and choline deficien ; MCDD, MCD die ; NASH, non-alcoholic
s ea ohepa i is; ROS, eac i e oxygen species; SC, egula die ; siRNA, small in e e ing RNA.
Jou nal o Hepa ology 2021 ol. 75 j34–45 39
hepa ocy es wi h unicamycin, an ER-s ess induce .
33
A dose-
dependen lipid accumula ion was obse ed a e 24 h o ea -
men , while siCnnm4 led o a dec ease in lipid con en (Fig. 5E
and Fig. S5B).
When cha ac e ising ROS p oduc ion in i o,siCnnm4- ea ed
mice showed educed dihyd oxie hidium (DHE) s aining (Fig.
5F), p obably as a esul o he oxida i e ac i i y measu ed by
a ious pa hways (Fig. S5C–F). Al hough oxida i e s ess was no
de e mined in he s udy wi h CD-HFD mice, he obse ed
educ ion in a y acid oxida ion (Fig. S5G) is consis en wi h a
ROS educ ion. The egula ion o he oxida i e s ess esponse
p omp ed us o e alua e ER s ess in he in i o models o NASH
using di e en ma ke s,
34
demons a ing up egula ion in bo h
NASH models and educ ion unde siCnnm4, sugges ing ha he
la e diminished he ER-s ess esponse (Fig. 5G and Fig. S5G).
ER-s ess associa ed signalling pa hways we e also e alua ed,
showing a dec eased ac i a ion o S6 ibosomal p o ein (S6) by
phospho yla ion.
35
CNNM4 inhibi ion inc eases MTP ac i i y, p omo ing VLDL
sec e ion
The o ma ion o p e-VLDL, he fi s s ep o VLDL assembly, akes
place in he ER.
36
Du ing his s ep, a small numbe o i-
glyce ides (TGs) a e associa ed wi h an apolipop o ein B (APOB)
molecule and embedded in a phospholipid monolaye by he
mic osomal iglyce ide ans e p o ein (MTP).
36
The al e a ions
o ER in eg i y p e iously obse ed unde MCD and he e e sion
o such by siCnnm4 sugges ed a possible modula ion o MTP
ac i i y. Rema kably, he p o ein was inc eased upon silencing
Cnnm4 in bo h he p ima y hepa ocy es (Fig. 6A) and in i o
oden NASH models (Fig. 6B).
The se um TGs o mice ed a 0.1% MCDD we e ound o
dec ease, owing o in ahepa ic lipid accumula ion, and we e
pa ially es o ed by Cnnm4 siRNA (Fig. S6A), sugges ing an
imp o ed VLDL sec e ion. The e o e, we de e mined he ela i e
APOB100 concen a ions in he se um, an indica o o ci cula ing
lipop o ein pa icles, obse ing an inc ease unde siCnnm4
F
BODIPY
siC lsiM p
MCD
siCnnm4C l siCnnm4C l
0
1
2
3
4
*
*
****
*
##
****
MCD + siCnnm4
C l
siCnnm4
MCD
siC l
siM p
In ensi y o lipid bodies/
cells ( old-change)
ΠΠΠΠ
ΠΠΠΠ
Π
E
BODIPY
VehicleLomi apide
MCD
siCnnm4siC lC l
0
2
4
6
8
**** ****
**** ****
###
MCD + siCnnm4
MCD + siC l
C l
In ensi y o lipid bodies/
cells ( old-change)
Vehicle
Lomi apide
ΠΠΠΠ
0
5
10
15
p = 0.1
D
0.1%MCDD + siCnnm4
0.1%MCDD + siC l
SC die
0-4 h
0-2 h
#
Li e TG sec e ion a e
(mM/kg/h)
*
#
C
SC die
CD-HFD + siCnnm4
CD-HFD + siC l
0.0
0.5
1.0
1.5
2.0
*
#
0.1%MCDD + siCnnm4
0.1%MCDD + siC l
SC die
0.0
0.5
1.0
1.5
2.0
Rela i e ApoB100 in
se um ( old-change)
B
0.0
0.5
1.0
1.5
2.0
*
#
SC die
CD-HFD + siCnnm4
CD-HFD + siC l
0.0
0.5
1.0
1.5
2.0 *
##
0.1%MCDD + siCnnm4
0.1%MCDD + siC l
SC die
Rela i e MTP ac i i y
( old-change)
A
0.0
0.5
1.0
1.5
2.0
2.5
****
####
MCD + siCnnm4
MCD + siC l
C l
Rela i e MTP ac i i y
( old-change)
Fig. 6. Inhibi ion o CNNM4 exp ession p omo es lipid expo by ac i a ing MTP ac i i y. (A) Rela i e MTP ac i i y in p ima y hepa ocy es cul u ed o 24 h in
a MCD medium and ans ec ed wi h a Cnnm4 siRNA (siCnnm4) o un ela ed con ol (siC l). (B) Rela i e MTP ac i i y in li e s om mice ed a 0.1% MCDD o a CD-
HFD and ea ed wi h siCnnm4 o siC l. (C) Rela i e apolipop o ein B100 (APOB100) le els in se a om mice ed a 0.1% MCDD o CD-HFD and ea ed wi h
siCnnm4 o siC l. (D) Li e TG sec e ion a e in mice ed a 0.1% MCDD, ea ed wi h siCnnm4 o siC l and a e poloxame P407 adminis a ion. BODIPY s aining
mic og aphs and quan ifica ion o p ima y hepa ocy es cul u ed unde MCD and ea ed wi h siC l o siCnnm4 and (E) a ehicle (DMSO) o 600 nM lomi apide
o 24 h o (F) an siRNA agains M p (siM p). Scale ba co esponds o 100
l
m. *p<0.05 and ****p<0.0001 s. SC die /C l + ehicle/C l + siC l;
#
p<0.05,
##
p<0.01,
###
p<0.001, and
####
p<0.0001 s. 0.1% MCDD + siC l/CD-HFD + siC l/MCD + ehicle + siC l/MCD + siC l;
p
p<0.05 and
pppp
p<0.0001 o ehicle/siC l s.
lomi apide/siM p. CD-HFD, choline-deficien high- a die ; MCD, me hionine and choline deficien ; MCDD, MCD die ; MTP, mic osomal iglyce ide ans e
p o ein; SC, egula die ; siRNA, small in e e ing RNA; TG, iacylglyce ide.
40 Jou nal o Hepa ology 2021 ol. 75 j34–45
Resea ch A icle NAFLD and Alcohol-Rela ed Li e Diseases
condi ions (Fig. 6C and Fig. S6B). The con ibu ion o inc eased
lipid sec e ion media ed by VLDL was u he confi med by
measu ing APOB100 in se um ex ac ed di ec ly om he ca a
ein, wi h VLDLs as he main componen , showing an inc eased
APOB100 con en (Fig. S6C). The hepa ic TG sec e ion a e, as
well as he VLDL con en , we e de e mined a e adminis e ing
poloxame P407 o inhibi lipop o ein lipase.
37
Rema kably, bo h
he sec e ion a e and he VLDL lipid con en ended o inc ease
upon silencing Cnnm4 (Fig. 6D and Fig. S6D). MTP was hen
inhibi ed in p ima y hepa ocy es by 24-h lomi apide incuba ion,
a selec i e inhibi o ,
38
o wi h a specific siRNA (siM p). As ex-
pec ed, he lipid educ ion obse ed upon CNNM4 knockdown
was absen unde lomi apide s imula ion (Fig. 6E) o M p
silencing condi ions (Fig. 6F and Fig S6E).
The possible a he ogenic seconda y e ec s a e a conce n
when i comes o p omo ing VLDL sec e ion. Fu he mo e, when
de e mining he Mg
2+
con en in isola ed VLDL, he obse ed
es o a ion in he g oup ea ed wi h he Cnnm4 siRNA (Fig. S7A)
sugges ed a possible impac on he WAT. The measu emen o
he ela i e a y acid oxida ion (FAO) capaci y o WAT e ealed a
endency o inc ease wi h he siRNA ea men (Fig. S7B),
oge he wi h an inc eased exp ession o genes in ol ed in lipid
oxida ion, mi ochond ial biogenesis, and he mogenesis (Fig.
S7C). This esul was co obo a ed in p ima y WAT adipocy es,
whe e Mg
2+
supplemen a ion inc eased glyce ol p oduc ion, an
indica o o lipolysis wi hou subsequen non-es e ified a y
acid (NEFA) p oduc ion (Fig. S7D). The same esul was obse ed
wi h condi ioned medium ob ained om p ima y hepa ocy es
cul u ed unde MCD and siCnnm4 condi ions (Fig. S7E).
The siRNA conjuga ion wi h GalNAc o e s a po en ial he apy
Conside able p og ess has ecen ly been made in he de elop-
men o oligonucleo ide-based he apeu ics, whe eby he
silencing o a ge genes is being explo ed in clinical s udies.
39
To
ans e he p esen ed disco e ies in o a he apeu ic app oach
sui able o clinical de elopmen , a newly iden ified Cnnm4
siRNA candida e was conjuga ed o an N-ace ylgalac osamine
(GalNAc) clus e ha binds o he asialo-glycop o ein ecep o s
p edominan ly exp essed in hepa ocy es. This echnology o e s
a po en ially sa e, specific, and e ficien mode o deli e y o
a ge ing he apeu ic molecules o hepa ocy es.
40
The specific inhibi ion o Cnnm4 was achie ed in p ima y
hepa ocy es h ough ecep o -media ed up ake by adding
di e en doses o a GalNAc-conjuga ed siRNA agains Cnnm4
(GalNAc siRNA) di ec ly o he cul u e medium (Fig. 7A). The lipid
accumula ion induced by MCD medium was educed by he
GalNAc siRNA ea men (Fig. 7B), and Mg
2+
accumula ion in
hepa ocy es was obse ed indi ec ly h ough i s dec ease in he
ex acellula medium (Fig. 7C). ROS o e p oduc ion upon MCD
s imula ion was also educed by ea ing he hepa ocy es wi h
he conjuga e (Fig. 7D).
Finally, he e ec i eness o Cnnm4 GalNAc siRNA was e alu-
a ed in i o in mice ed a 0.1% MCDD o 6 weeks, leading o a
mo e se e e NASH pheno ype.
21
The ea men was ini ia ed a
he 3-week poin wi h a single injec ion o di e en doses (1 o 5
mg/kg) o Cnnm4 GalNAc siRNA (0.1% MCDD + GalNAc siRNA) o
a con ol siRNA conjuga e (0.1% MCDD + siC l). A specific and
significan inhibi ion o Cnnm4 mRNA exp ession was de ec ed in
mice adminis e ed ei he 1 o 5 mg/kg (Fig. 7E), while he Mg
2+
le els in se um we e educed in line wi h a p esumed hepa ic
Mg
2+
accumula ion (Fig. 7F). Impo an ly, CNNM4 inhibi ion by
GalNAc siRNA significan ly educed hepa ic lipid accumula ion
and alle ia ed he inflamma o y esponse and fib osis de elop-
men (Fig. 7G).
Discussion
The s udies desc ibed he ein pinpoin CNNM4 as a key egula o
o Mg
2+
homeos asis in hepa ocy es and a po en ial he apeu ic
a ge o NASH. To da e, esea ch pe o med on his ca ion in
li e pa hologies has shown a p o ec i e e ec o Mg
2+
supple-
men a ion
4
and e ealed Mg
2+
deficiencies in pa ien s wi h
ci hosis o li e cance .
18
Indeed, hypomagnesaemia is
equen ly obse ed in NASH como bidi ies.
6,7
Howe e , he e is
s ill a dea h o knowledge on he implica ions o Mg
2+
pe u -
ba ions o he de elopmen o NASH, and no hing has been
epo ed abou magnesio opic p o eins and hei modula ion in
he li e .
P e ious s udies ha e mainly ocused on PRL, he in e ac ing
pa ne o CNNM, associa ing i s exp ession wi h poo p ognosis
in mul iple cance ypes including li e ,
16,17
whe eas s udies o
CNNMs ha e been limi ed o hei ole in anspo ing Mg
2+
ac oss epi helia.
25
He ein, we demons a e ha CNNM4 is a
con ibu o o NASH, as i is o e exp essed in p eclinical models
and clinical samples o he pa hology. The educ ion in CNNM4
exp ession in THLE-2 cells upon inhibi ing NEDDyla ion sugges s
he in ol emen o pos - ansla ional mechanisms in he mod-
ula ion o he s abili y o CNNM4 du ing NASH. This is u he
confi med by he educ ion o MCD-induced lipid accumula ion
by he specific silencing o Cnnm4, which does no happen upon
silencing o he Cnnms. Rega ding he e ec o a ge ing Cnnm4
on NASH de elopmen , p eclinical s udies in oden models
showed p omising e ec s o an siRNA-based he apy in mice ed
ei he a 0.1% MCDD o CD-HFD. Significan ly, bo h die s inc ease
lipid accumula ion and fib osis de elopmen , bu he siRNA
he apy amelio a es bo h hallma ks o he disease. The ole ha
CNNM4 may play in o he cell popula ions such as hepa ic
s ella e cells equi es mo e esea ch, as do he possible pos -
ansc ip ional and pos - ansla ional mechanisms ha may
con ibu e o CNNM4 o e exp ession.
The ole o he hepa ocy e in NASH and i s subsequen p o-
g ession has been widely cha ac e ised.
27,28
The silencing o
Cnnm4 may educe oxida i e ac i i y, leading o educed oxida-
i e s ess, which could con ibu e o a educ ion in fib osis
de elopmen . Al hough he de elopmen o oxida i e s ess was
no cha ac e ised in he CD-HFD in i o s udy, he obse ed
educ ion in FAO and lipid con en s p omp s o expec a simila
educ ion. The ele ance o he hepa ocy e in NASH p og ession
and he implica ions o modula ing CNNM4 we e u he cha -
ac e ised in mice ed a 0.1% MCDD o 6 weeks. He ein, we also
demons a ed ha a ge ing CNNM4 amelio a ed NASH, e en
ha wi h a mo e se e e pheno ype.
21
To de elop a possible
he apeu ic app oach, an siRNA was conjuga ed wi h GalNAc
allowing s able and li e -specific deli e y,
40
wi h esul s simila
o hose obse ed when a ge ing Cnnm4 wi h he liposomal
siRNA o mula ion a e 4 weeks on he 0.1% MCDD.
Al hough CNNM4 has been cha ac e ised as a Mg
2+
ans-
po e in kidney epi helia,
41
he CNNM4-media ed flux o he
ca ion in he li e emains unknown. Ou esea ch shows
CNNM4 o be a Mg
2+
expo e in he hepa ocy e, as i s specific
silencing inc eases in acellula Mg
2+
. Hepa ic siCnnm4-induced
Mg
2+
accumula ion was cha ac e ised in all he p eclinical
s udies h ough he quan ifica ion o he ca ion in se um,
Jou nal o Hepa ology 2021 ol. 75 j34–45 41
